Bruce Momjian <bruce(at)momjian(dot)us> writes:
> For me, 'struct timeval' is coming in via #include "libpq/libpq.h", but
> of course that is _after_ the inclusion of resource.h.

Ah, that explains it.

> Not sure where
> you see that sys/resource.h was always there.  Looking at the CVS diffs
> I see it added only in the past day.

It was further down before.

There's still something pretty strange here, though, because AFAICS
configure should have rejected sys/resource.h if it needs sys/time.h.

			regards, tom lane

In response to

• Re: [COMMITTERS] pgsql: Add include needed for new at 2006-10-08 18:14:03 from Bruce Momjian

Responses

• Re: [COMMITTERS] pgsql: Add include needed for new at 2006-10-08 21:32:14 from Bruce Momjian
• Re: [COMMITTERS] pgsql: Add include needed for new getrusage() call. at 2006-10-08 21:47:04 from Peter Eisentraut

pgsql-hackers by date

Next:	From: Chris Campbell	Date: 2006-10-08 18:43:28
	Subject: Re: [PATCHES] Use non-deprecated APIs for dynloader/darwin.c
Previous:	From: Tom Lane	Date: 2006-10-08 18:29:25
	Subject: Re: Use non-deprecated APIs for dynloader/darwin.c

pgsql-committers by date

Next:	From: Bruce Momjian	Date: 2006-10-08 19:18:29
	Subject: Re: pgsql: In release notes, mention that
Previous:	From: Jaime Casanova	Date: 2006-10-08 18:26:14
	Subject: Re: pgsql: In release notes, mention that multi-line history saving does not